Developing Country Fiscal Policy Responses to the Crisis
Completion Reports
Knowledge for Change Program (KCP)
Summary Information
TF095859
Developing Country Fiscal Policy Responses to the Crisis
KCPII Window
Investment Climate & Trade and Integration
Country
World
TTL Name
Aart Kraay
Project Period
5/16/2009 – 6/30/2011
Approving Manager
Luis Serven
Grant Amount
50,000.00
Disbursement
44,940.00
Â
Â
Â
Research Team
Aart Kraay, World Bank, Research Department Luis Serven, World Bank, Research Department Ethan Ilzetzki, London School of Economics
Research Objective
1. Summary of Aims, Objectives and Significant Achievements
Objective:Â
The global crisis has put under the spotlight the ability of developing countries to resort to fiscal policy to counteract the deepening recession. A major difficulty in designing appropriate fiscal policy responses in developing countries is the near total lack of systematic empirical evidence on the actual effects of fiscal policy changes on output. This proposal financed original research designed to isolate the causal effects of fiscal policy on growth in the short and long run in developing countries. The objective was to inform two central questions in the design of fiscal policy responses. How large should they be to have the desired short-term impact? Will they be sustainable in the long run?
Significant achievements:Â
As detailed below, the research project has led to the creation of a new dataset of marginal tax rates in developing countries, and two new papers deploying innovative approaches to understanding the short- and long-run effects of fiscal policy on output. The papers have been disseminated through the Bank's working paper series, through seminars, and through submission for journal publication.
2. Nominated Outputs
Output 1: Kraay, Aart (2010) "How Large is the Government Spending Multiplier? Evidence from World Bank Lending". World Bank Policy Research Department Working Paper No. 5500. Available at: http://go.worldbank.org/MACZF9XY50 This paper proposes a novel approach to empirically identifying government spending multipliers, that relies on two features unique to many low-income countries: (1) borrowing from the World Bank finances a substantial fraction of government spending, and (2) spending on World Bank-financed projects is typically spread out over several years following the original approval of the project. The first fact means that fluctuations in spending on World Bank-financed projects are a significant source of fluctuations in overall government spending in these countries. The second fact means that fluctuations in World Bank-financed spending in a given year are largely determined by fluctuations in project approval decisions made in previous years, and so are unlikely to be correlated with shocks to output in the current year. The paper uses World Bank project-level disbursement data to isolate the component of World Bank-financed government spending in a given year that is associated with past project approval decisions. This can then be used as an instrument for total government spending, in order to estimate spending multipliers in a sample of 29 mostly low-income countries where this source of variation in government spending is large relative to the size of the economy. The resulting spending multipliers are small, reasonably precisely estimated, and insignificantly different from zero in most specifications.
Output 2: Ilzetzki, Ethan (2011). "Fiscal Policy and Debt Dynamics in Developing Countries". World Bank Policy Research Working Paper No. 5666. Available at: http://go.worldbank.org/NPQSA1AXQ0. Using a new tax database for 28 countries and a variety of econometric methods, this paper contributes to the debate on the effects of fiscal policy on economic activity in a number of ways. The analysis finds that tax cuts have a stimulative effect on economic growth in developing countries. Lowering the personal income tax rate by 1 percentage point, or cutting revenues by 1 gross domestic product of gross domestic product increases gross domestic product by 0.3-0.4 percent on impact and 0.8 percent in the long run. The author finds that cuts in personal income taxes are more effective in stimulating growth than cuts in corporate or valued added tax rates. The author incorporates debt dynamics into a fiscal vector autoregression model for a number of developing countries. Existing estimates of the effects of fiscal policy on growth use linear time-series methods, which may assess the effects of fiscal policy along a debt-path that is unsustainable. Incorporating the non-linear relationship between government expenditure, taxes, and debt alters estimates of the impact of fiscal policy on gross domestic product in several countries. In Brazil, for example, conventional time-series methods may overstate the effects of fiscal policy on gross domestic product, by ignoring the detrimental effects of debt accumulation.
Output 3: Dataset on aggregate effective tax rates in developing countries. This dataset combines statutory data on various tax rates, together with macroeconomic and household survey data on the evolution of various tax bases, to construct a panel dataset of effective tax rates cover 28 countries, including 19 developing countries. Available at: http://personal.lse.ac.uk/ilzetzki/data.htm.
3. Major Difficulties
No major difficulties were experienced in carrying out the research envisioned in the original proposal.
4. Surprises
One surprise (relative to the original proposal) was the development of an entirely new methodology to identify exogenous shocks to government spending that could be used to identify government spending multipliers (as detailed in nominated output 1).
A second surprise was how well very basic methodologies for estimating aggregate effective tax rates using the limited available data for developing countries seem to work. As described in nominated output 2, applying the very basic assumptions used to calculate effective tax rates in developing countries to US data resulted in very similar estimates of tax rates to those obtained by other authors using far more sophisticated information available for the US.
5. Full Completion Report
• Background and motivation
The global crisis has put under the spotlight the ability of developing countries to resort to fiscal policy to counteract the deepening recession. A major difficulty in designing appropriate fiscal policy responses in developing countries is the near total lack of systematic empirical evidence on the actual effects of fiscal policy changes on output. This proposal financed original research designed to isolate the causal effects of fiscal policy on growth in the short and long run in developing countries. The outcome of this research will inform two central questions in the design of fiscal policy responses. (1) How large should they be to have the desired short-term impact? and (2) Will they be sustainable in the long run? Answers to these questions are of great interest to the Bank as it currently struggles to provide advice to developing country policymakers on the appropriate fiscal response to the crisis. The fact that, even for the United States where, despite an abundance of high-quality empirical work there is little consensus as to the appropriate size of a fiscal stimulus and its longer-term consequences for debt, highlights the importance of careful empirical work in this area for developing countries to inform such policy discussions.
• Original objectives of the research, along with any changes and their justification; ; note their consistency with Bank, DECRG, and KCP objectives
The original objectives of the research were to generate 2-3 research papers, together with supporting data, to provide new insight into the questions noted above.
• Assessment of the extent to which the objectives have been met, with explanations of any shortfall.
These objectives have broadly been met. The three nominated outputs described above correspond to the basic initial goals. One modification relative to the original objectives was a change in focus in one of the papers. The original proposal envisioned a separate paper on the longer-run growth impacts of productive public expenditures. This paper was replaced with a paper on estimating short-run spending multipliers in low-income countries, where such analysis had not been carried out before.
• Methodology and data
This project has extended methodological techniques and gathered new data required to analyze the short- and long-run effects of fiscal policy on output. The fundamental identification problem in this literature is straightforward: in order to identify the effects of government spending changes, or tax changes, on output, it is necessary to isolate the component of such changes that is plausibly uncorrelated with contemporaneous macroeconomic events. The first nominated output develops an entirely new approach to doing so, based on disbursement lags in World Bank financed project. A key feature of such projects is that they take a long time to implement. As a result, in any given country-year, actual World Bank disbursements are in large part determined by project approval decisions made in previous years, and before current shocks are known. In low-income countries where World Bank lending finances a large share of public spending, this insight can be used to isolate fluctuations in public spending that are predetermined by these past project approval decisions, and so are likely to be uncorrelated with current shocks. The paper uses this to construct an instrument for changes in total public spending, and finds that these have at best small and statistically insignificant effects on output in the short run. The second paper relies on a more standard assumption from the literature to achieve identification: that fiscal policy responds sufficiently slowly that quarterly changes in spending or taxes are unlikely to be correlated with contemporaneous quarterly shocks. Using quarterly data in a sample of emerging economies, it first estimates government spending multipliers based on this identification approach. It then extends both the data, and the type of analysis, in two important ways. First, as described above, it constructs estimates of aggregate effective tax rates for a sample of 28 countries (including 19 developing countries), that did not previously exist. It then uses both the tax and spending data to trace out the longer-term effects of changes in both, also taking into account the implied path of future debt accumulation. This point has been ignored in most of the empirical literature on estimating fiscal multipliers, and this paper is the first to apply such techniques to developing country data.
• Results: What have we learnt so far?
The main finding from the first nominated output is that, in a sample of 29 aid-dependent low-income countries, the short-run effects of government spending on output are small, and generally statistically insignificant. The paper estimates that an additional dollar of government spending raises output on average by about 50 cents, indicating a very substantial degree of crowding out of private economic activity by increases in government spending. While this finding casts doubt on the effectiveness of government spending as a countercyclical tool in these countries, this does not imply that there is no role for public spending in response to adverse economic shocks: there is for example considerable scope for expanding social welfare programs during economic downturns. However such interventions are best thought of as providing social protection rather than as providing Keynesian stimulus. The main message from the second nominated output is the crucial importance of taking into account paths of future debt accumulation when assessing the short-run effects of government spending and tax changes. The paper reports evidence from Brazil suggesting that the failure to take debt accumulation, and resulting interest rate changes, into account can significantly overstate the short-run stimulative effects of spending increases and tax cuts.
• Dissemination details (including future plans)
The resulting papers will be disseminated through the working paper series and by journal submissions, as well as conference presentations and through the web. The target audience consists of policymakers and researchers.
• Impact—within and outside the Bank
The results from this research project have been widely-disseminated inside and outside of the Bank. It is hoped that the cautionary note struck by both of them on the effectiveness of short-run fiscal stimulus packages will contribute to introducing some more caution in policy discussions of the topic.
• Implications for future research
The new approach to identifying exogenous fluctuations in government spending based on disbursement lags in aid-financed spending projects can be applied to a range of other interesting questions, including the additionality of foreign aid, and as a way of developing more plausible instruments that can be used to analyze the longer-term growth effects of foreign aid.
• Capacity building
This project has no direct capacity-building objectives. One result this project has contributed to is to demonstrate the feasibility in a developing-country contextof using state-of-the-art techniques used to identify fiscal policy shocks and their consequences in industrial countries
• Additional resources leveraged
None
• Operationalization and development policy implications
As discussed above, the results in these papers have important implications for designing fiscal policy responses to macroeconomic shocks in the first run. The main message from the first is that policymakers should contemplate fiscal stimulus with a strong awareness of the fact that short-run output responses are likely to be small. The main message from the second is that policymakers should carefully consider the adverse effects of debt accumulation driven by deficit-financed fiscal stimulus.
Annex 1: Reference list and background papers
Kraay, Aart (2010) "How Large is the Government Spending Multiplier? Evidence from World Bank Lending". World Bank Policy Research Department Working Paper No. 5500. Ilzetzki, Ethan (2011). "Fiscal Policy and Debt Dynamics in Developing Countries". World Bank Policy Research Working Paper No. 5666.